Missing Girls in Spain Libertad González May 2014
Total Page:16
File Type:pdf, Size:1020Kb
Missing Girls in Spain Libertad González May 2014 Barcelona GSE Working Paper Series Working Paper nº 760 Missing girls in Spain Libertad González (Universitat Pompeu Fabra and Barcelona GSE) May 2014 Abstract: I document extremely son-biased sex ratios at birth among Asian-born parents in Spain. Using data on the universe of registered births in Spain during 2007- 2012, I show that there are 117 boys per 100 girls born to Indian parents, compared with a national average of 107 boys per 100 girls. The difference is even more pronounced at higher parities. Indian parents with one or more previous children have 143 boys per 100 girls (117 for Chinese parents with at least two previous children). These biases are significant even after controlling for a broad range of family characteristics. The most likely mechanism is sex-selective abortions. I find no increase in the sex ratio after the 2010 reform that deregulated abortion during the first 14 weeks of pregnancy. JEL codes: J13 Keywords: fertility, sex ratio, abortion, sex selection, son preference, India, Spain. I thank one anonymous referee and the students in my graduate Labor Economics class at the Barcelona GSE in the Spring of 2014 for their useful comments. Any remaining errors are my own. 1. Introduction It is well-known that fertility rates are higher among immigrants than natives, in Spain as well as in other countries (Adsera and Ferrer, 2013). For instance, in 2012 the birth rate among Spanish nationals was about 44 births per 1,000 women, compared with 51 for foreign nationals (see Figure 1).1 It has also been documented that the fertility patterns of immigrants tend to reflect those in the source country (Fernández and Fogli, 2009), although there is also evidence of assimilation in fertility behaviors (Mayer and Riphahn, 2000). Several Asian countries (notably India and China) are known for their increasingly biased sex-ratios in favor of boys (Jha et al., 2006, Abrevaya, 2009), a phenomenon which is referred to by the term “missing girls”. There is some recent evidence that these patterns are also present among Asian immigrants in developed countries (see Abrevaya, 2009 and Almond & Edlund, 2008 for the US, Dubuc and Coleman, 2007 for the UK and Almond et al., 2013 for Canada). In this paper I use birth-certificate data to document for the first time son-biased sex-ratios at birth among Asian immigrants in Spain. This bias is particularly pronounced among parents born in India, and for higher parities, and it remains significant after controlling for a wide range of family characteristics. I argue that the most likely mechanism driving the low fraction of newborn girls is the practice of sex-selective abortions among (some) Indian families in Spain.2 Other methods potentially used by families with a preference for sons, such as selective fertility stopping (families that stop having children as soon as they have a son), cannot 1 Source: Spanish National Statistical Institute (www.ine.es). 2 Other methods for prenatal sex-selection, such as gender-selective in-vitro fertilization and sperm sorting, are illegal for non-therapeutic reasons in Spain. 1 by themselves generate a biased sex ratio at birth (although they would generate other patterns such as girls having on average more siblings than boys). In 2010, a reform in the legal regulation of abortion in Spain made it easier to get an abortion during the first 14 weeks of pregnancy. I evaluate the effect of this reform on the degree of sex-selection among Asian parents, but find little evidence that the reform led to any increase in the prevalence of sex-selection. This is consistent with the most widely used methods for sex determination (ultrasound and amniocentesis) not being available before week 16 of the pregnancy. The remainder of the paper is organized as follows. Section 2 introduces the methodology and the data used in the analysis. Section 3 presents the results, and section 4 concludes. 2. Methods and data The main data source comes from birth-certificate data, as made publicly available by the Spanish National Statistical Institute (www.ine.es). The micro-data files provide information on the universe of births registered annually in Spain, including some demographic variables for the parents. I focus on years 2007-2012, since before 2007 no information was provided on the country of birth of the parents. As an additional data source, I also explore the aggregate data provided by the National Statistical Institute on registered abortions (by weeks of gestation, country of birth of the mother, etc). I focus on singleton births (less than 2% of all births were multiple). There were a total of 2,810,143 registered singleton births in Spain between 2007 and 2012, almost 18% of them with both parents born outside of Spain (see Table 1). Almost 8% of newborns with foreign-born parents (about 38,000) had both parents born in Asia. There are 24,499 births with both parents born in China (26,977 with at least one parent born 2 in China), and 2,805 with both parents born in India (3,665 with at least one parent born in India).3 The descriptive analysis starts by calculating sex ratios at birth. The sex ratio is defined as the number of boys born per 100 girls. A sex ratio of about 105 is considered “biologically normal” (Almond & Edlund, 2008). The econometric analysis estimates the following equation on the sample of all singleton births: (1) Bic = + Iic + Xic + ic Each observation refers to an individual newborn baby i, with parents born in country c. The dependent variable takes value 1 if the newborn is a boy (0 if it’s a girl). The main explanatory variable, I, is an indicator that takes value 1 if both parents are born outside of Spain. A second specification defines I as equal to 1 if both parents are Asian-born, and a third one uses a set of indicators for specific source countries. I include control variables X that might be related to the sex of the baby (age of the mother and father, education of the mother, year dummies, and number of previous children). The coefficient of interest, , will be different from zero if immigrant parents are more (or less) likely than natives to give birth to boys. In additional specifications, I interact the indicators of country of birth of the parents with birth order. As has been noted in the literature (Abrevaya 2009), once a family approaches its maximum size (the desired total number of children), the incentives for sex selection increase if there is a strong preference for one sex over the other (given the sex composition of the previous children). Thus, we would expect that 3 There are 7 Asian countries reported separately in the birth-certificate data. The Asian country with the highest number of immigrant parents is China (65%), followed by Pakistan (15%), the Philippines (8%) and India (7%). 3 parents with a preference for boys will be more likely to have sex-selective abortions, the higher the birth order of the child. Ideally, we would like to be able to control for the sex of previous children born to the same parents. We would expect that, among parents with a preference for boys having a third child, a family where the first two children were girls will be more likely to select the sex of the third child, compared with a family with two boys. However, the Spanish birth-certificate data do not provide information on the sex of the previous children. 3. Results 3.1 Descriptive results Descriptive statistics for the sample of singletons born in Spain in 2007-2012 are presented in Table 1. The fraction of boys among all singleton births was 51.7%, which translates into an overall sex ratio at birth of 106.9 boys born per 100 girls ((51.7x100)/48.3), with no particular trend apparent over the time period analyzed. Mothers are on average 31 years old at birth, while fathers are about two years older. Almost 2% of babies have no registered father. About 27% of the mothers have a high- school degree but no college, and 30% are university-educated. Regarding birth order, 54% of births are first-born children for the mother, while 36% are second-born, and only 10% of mothers had 2 or more previous children. Figure 2 shows the sex ratio at birth for native parents, all immigrant parents, and Chinese and Indian parents. The sex ratio is not significantly different between native (106.8) and foreign-born parents (107). The sex ratio for Chinese parents is also very close to the one for natives (107.1). However, Indian parents in Spain are significantly less likely to give birth to girls, compared with native parents as well as with other 4 immigrant source countries. Indian parents had 117 boys per 100 girls. This suggests that the practice of sex-selective abortions, documented in India, might also be prevalent among Indian families residing in Spain. Figure 3 breaks down the sex ratios by birth order, showing the number of boys born per 100 girls, separately for women having their first, second, third and higher- order child. The sex ratio in the full sample is close to 107 for all birth orders. Chinese parents have 106 boys per 100 girls in first and second births. However, third births to Chinese parents show a sex ratio of 116, and the ratio is 122 for higher-order births. This suggests some degree of sex selection among Chinese parents with two or more previous children.4 The trend becomes striking in the case of Indian parents.